Rent and Run

The Rent and Run module within the Grafi Cloud Console allows renters to easily access and utilize GPU resources available on the Grafilab network. Whether you need computational power for gaming, AI model training, machine learning research, or rendering, Rent and Run provides a seamless way to select, rent, and execute tasks using both centralized and decentralized GPUs.

How Rent and Run Works

  1. Search for Available GPU:

    • Through the Grafi Cloud Console, renters can search for available GPUs based on your requirements, such as GPU type, task duration, and pricing.

    • The available nodes are categorized into:

      • CEPIN (Centralized Physical Infrastructure Network): GPUs from data centers or cloud providers like AWS, Alicloud, or partner data centers.

      • DEPIN (Decentralized Physical Infrastructure Network): GPU resources provided by individual contributors using the Grafi Agent.

  2. Select GPUs:

    • Choose from a list of available GPU based on performance specifications, cost, and availability. Each GPU will have details like GPU model, memory, and estimated processing power.

  3. Run Your Task:

    • Once you’ve selected your preferred GPU, you can SSH into the virtual machine and set up your application.

    • The virtual machine is executed in a Docker container, ensuring a standardized environment across different GPUs.

  4. Real-Time Monitoring:

    After deploying the virtual machine, you can monitor its progress in real-time through the Grafi Cloud Console. The console displays instance status, GPU performance metrics, and estimated cost.

  5. Data Aggregation:

    • All events happen across the GPU to rental subscription are sent to Grafi AI node for indexing and push to data layer via smart contracts, ensuring transparency in compute power, payments and task execution.

Last updated